A survey and taxonomy of loss functions in machine learning
This is an incremental survey that provides a taxonomy for machine learning practitioners and students to better understand loss functions.
The paper tackles the problem of understanding and selecting loss functions in machine learning by providing a comprehensive survey and taxonomy of 43 distinct loss functions across key applications, resulting in a structured resource for students and researchers.
Most state-of-the-art machine learning techniques revolve around the optimisation of loss functions. Defining appropriate loss functions is therefore critical to successfully solving problems in this field. In this survey, we present a comprehensive overview of the most widely used loss functions across key applications, including regression, classification, generative modeling, ranking, and energy-based modeling. We introduce 43 distinct loss functions, structured within an intuitive taxonomy that clarifies their theoretical foundations, properties, and optimal application contexts. This survey is intended as a resource for undergraduate, graduate, and Ph.D. students, as well as researchers seeking a deeper understanding of loss functions.